A Big Data Analytics Framework for Supporting Multidimensional Mining over Big Healthcare Data

Abstract

Nowadays, a great deal of attention is being devoted to big data analytics in complex healthcare environments. Fetal growth curves, which are a classical case of big healthcare data, are used in prenatal medicine to early detect potential fetal growth problems, estimate the perinatal outcome and promptly treat possible complications. However, the currently adopted curves and the related diagnostic techniques have been criticized because of their poor precision. New techniques, based on the idea of customized growth curves, have been proposed in literature. In this perspective, the problem of building customized or personalized fetal growth curves by means of big data techniques is discussed in this paper. The proposed framework introduces the idea of summarizing the massive amounts of (input) big data via multidimensional views on top of which well-known Data Mining methods like clustering and classification are applied. This overall defines a multidimensional mining approach, targeted to complex healthcare environments. A preliminary analysis on the effectiveness of the framework is also proposed.